Skip to content

AI infrastructurein a glass boxyou own

Build and ship AI applications on infrastructure you own — your hosts, your keys, your choice of model.

$git clone https://github.com/digithings-ai/digithings.git

Ask digichatScroll to explore
// by the numbers

The platform, in four numbers.

No asterisks — every figure is countable in the repo. Nine modules ship today and two are still on the roadmap; the manifest below names which.

modules shipping
compose services
vector backends
keys stored
// the architecture

Nine modules. One toolkit.

The nine that ship run standalone or compose with the rest — retrieval, quant research and chat, plus the auth, tracing and audit any deployment needs. Two more are on the roadmap; the manifest below marks which.

// modules · 9 online · 2 on the roadmap

select a module

// integrations

Built on what you already run.

Every module is assembled from open-source libraries you can name, version, and swap — no forks, no reimplementations. That is why this drops into an existing deployment instead of asking you to replace one.

orchestration & models
LangGraphLiteLLMCheaper InferenceMCPOpenAI SDKPydanticFastAPI
quant, data & retrieval
NautilusTraderOptunaPolarsChromaAzure AI SearchSupabasePostgres
runtime, surface & observability
DockerOpenTelemetryPrometheusRedisNext.jsReactAssistant UI
// the repository

Maintained in the open.

One MIT-licensed monorepo, public and moving. Below are the most recent pull requests to land on main — each one you can open and read, not a changelog entry we wrote about ourselves.

// last 30 days on main

  • 2,388commits
  • 1,126PRs merged
  • 532issues closed

// current backlog

  • 2PRs open
  • 163issues open

// contributions — last year

8,758 contributions in the last year

latest digichat-v1.5.0

snapshot 2026-09-15

browse the repo
git clone https://github.com/digithings-ai/digithings

// merged recently

  1. #4094fix(website): keep repo-activity openIssues to real issues (#4091)2026-09-15
  2. #4090chore(promote): module/digiquant -> develop (#4069)2026-09-15
  3. #4092chore(sync): module/digiquant <- develop2026-09-15
  4. #4088chore(release): promote develop to main — web-grounding retry + R2-only market reads2026-09-15
  5. #4085feat(digiquant): digifetch x Gloomberb market-data tools (#4069)2026-09-15
  6. #4087fix(digiquant): retry web grounding unscoped when a scoped search is empty (#4086)2026-09-15

// open issues

  1. #4101ops(digiquant): post-deploy verification of digifetch_* tools + upstream drift checks2026-09-15
  2. #4100feat(digiquant): widen 1wk history range via startDate (upstream serves >5Y)2026-09-15
  3. #4099docs(digiquant): runbook for GLOOMBERB_SESSION_COOKIE (free account + per-deployment env)2026-09-15
  4. #4098feat(digiquant): surface Gloomberb deep links + attribution in digiquant pages and digichat2026-09-15
  5. #4097feat(digiquant): wire /v1/orchestrator_invoke dispatch for the digifetch_* tool family2026-09-15
  6. #4096ops(digiquant): verify Gloomberb MCP image builds + in-container egress (last #4069 spike)2026-09-15
// why digithings

Nobody knows which vendor wins.

The field moves faster than any bet you could place on it, so this architecture declines to place one. Nothing here is married to a provider: the stack is self-hosted, keys arrive per request, and models and vector backends are configuration rather than code. When the leader changes, you change one setting — not the application on top.

  1. Self-hosted by default

    One docker-compose file, sixteen services, on a laptop or any host you own — eight up by default and the rest behind named profiles you turn on: digichat, digivault, the search MCP, the heartbeat, observability, and the LiteLLM cache.

  2. BYOK, every request

    Anthropic, OpenAI, or any LiteLLM-compatible key — forwarded per-request, never stored.

  3. Audit-on by default

    Append-only JSONL audit and a correlation ID on every service hop. Events record a prompt's length, never its text — tail events.jsonl and check.

  4. Backend-swappable

    Two live vector backends — Chroma and Azure AI Search — behind one client, swappable without touching business code.

Contact

Questions, enterprise, or partnership.

The whole monorepo is MIT-licensed and public — take it and run it yourself. What we sell is the integration work: fitting these modules to the stack you already have, on your own infrastructure.

Or email us directly